Application 3D Forensic Science in a Criminal Investigation

IRENA DROFOVA, MILAN ADAMEK, PAVEL STOKLASEK, MARTIN FICEK,

JAN VALOUCH

Faculty of Applied Informatics,

Tomas Bata University in Zlin,

Nad Stranemi 4511, Zlin 760 05,

CZECH REPUBLIC

Abstract: - This manuscript discusses the modern approach and application of 3D digital imaging in forensic

science. It presents the basic principles and approaches of 3D modeling methods. Selected methods of image

capture and its subsequent processing into a 3D model are applied to a specific object. This object is captured

by a mobile phone camera, a LiDar sensor, and a 3D scanner for further image processing for different desired

image outputs. The text describes the photogrammetry method, the workflow with the LiDar sensor, and the 3D

model of the object intended for 3D printing. The paper discusses the potential of the selected methods and

their application in forensic sciences.

Key-Words: - forensic science, criminal investigation, 3D model, photogrammetry, LiDar sensor, 3D print

Received: April 28, 2022. Revised: January 12, 2023. Accepted: February 7, 2023. Published: March 2, 2023.

1 Introduction

Modern digital methods of displaying objects and

scenes find application in many fields. Digital image

data capture reality or faithfully create a new reality.

3D modeling methods are currently commonly

encountered in many commercial and scientific

fields. 3D visualization and virtual and augmented

reality (VR/AR) are used today in music, [1],

architecture, [2], art, [3], medicine, [4], forensics

science, [5], advertising, military, industrial design,

and agriculture, for capturing and digitizing cultural

heritage and forgeries. Many methods of 3D

modeling and visualization find their application

across the fields of human activity.

Modern procedures and methods of 2D and 3D

visualization are also applied in criminology and

forensic sciences, [6]. Currently, 3D reconstructions

of objects are used in courts, [7]. Investigators use

them to support evidence from the crime scene and

are also used in expert opinions. Thus, a specific

branch of forensic science emerged: 3D forensic

science (3DFS), [8]. This new field provides new

perspectives in forensic art and technology. 3D

scanning methods and visualization can faithfully

reconstruct the appearance of the victim of a crime,

the crime scene, and the evidence found at that

location, [9]. 3DFS can also be used for the

reconstruction of historical image 2D data. It is

essential to mention that this is a virtual

representation of the data; therefore, the outputs

from the 3DFS methods are only supplementary in

case of investigation. Nevertheless, the use of 3DFS

methods is also in great demand in court

proceedings. The created 3D objects are used for

display on court monitors and for creating a

database of evidence in criminal proceedings, [8].

The issue of 3DFS connects many fields. These

are mainly computer graphics, computer vision,

machine learning, image capture and display, and

other scientific disciplines. The output can be a

virtual digital image for display on screens or in

virtual and augmented reality. Image data can be

used for 3D printing, [10]. Selective image capture

and image data processing procedures are chosen

depending on the type of output. Individual

procedures differ slightly, especially for the display

of material and its accurate, realistic display.

Reconstruction of metallic materials with a high

gloss level tends to be problematic. Structured

materials also often require an individual approach

when processing image data, [11]. Therefore, when

investigating a crime and securing evidence using

3DFS, emphasis is placed on optimizing the entire

process and speed of managing input image data.

This manuscript presents selected basic methods

of 3D reconstruction of a real object for 3D

visualization and obtaining a virtual model for 3D

printing purposes. The experiment uses the method

of photogrammetry and 3D scanning. The

experiment's goal is to rapidly process image data

into a 3D model for further use based on the

required outputs. A smartphone is used in this

experiment. These devices are currently equipped

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2023.20.8

Irena Drofova, Milan Adamek,

Pavel Stoklasek, Martin Ficek,

Jan Valouch

E-ISSN: 2224-3402

Volume 20, 2023

with a standard camera for obtaining high-quality

photos. The object scanned by the smartphone was

transformed into a 3D model using the

photogrammetry method. Next, a LiDar sensor was

used to capture the same object. This sensor was

part of only highly professional and costly devices

in the past. Currently, it has reached the higher class

of tablets for ordinary users. These two devices can

replace classic cameras and streamline the work of

technicians and processing image data from the

crime scene. The object was then scanned with a

scanner for 3D printing output. Here, high and

accurate quality of the reproduction of the details of

the object's defects is assumed. This work aims to

demonstrate the contribution of new modern devices

and their application in criminalistics. Especially at

the crime scene as a tool to secure evidence and

quickly transform it into a digital 3D environment.

2 3DFS Methods and Procedures

This chapter presents the basic 3DFS methods and

procedures for obtaining a 3D model of an object.

The object for 3D reconstruction is the spent charge.

The object is made of plastic and metal, with a relief

marking the charge. The experimentally created

models are presented in the following chapters.

2.1 Image Capture

Several imaging devices were used in this

experiment. The following devices were used to

obtain a digital image for the subsequent

reconstruction of the object using the

photogrammetry method:

• Cubot X30 mobile phone

• iPad 11 Pro with LiDar sensor

To 3D model for subsequent output in the model to

3D printing, the following equipment was used:

• 3D scanner Atos Triplescan II 5M / GOM

It is clear from the listed image capture devices that

these are three different devices. In the case of a

mobile phone and tablet with a LiDar scanning

function, it is possible to easily create data for

output through 3D printing. Acquiring a 3D model

with an Atos scanner and transforming it into a 3D

color model is complicated by the absence of color

in the base 3D model.

2.1.1 Photogrammetry Method

This method was applied and perfected already at

the time of analog image processing. With the

development of digitization, this method finds other,

more comprehensive applications. That is a method

that is not demanding in the requirements of the

sensing device.

In this experiment, the method of multi-frame

ground photogrammetry was used. The principle of

this method is to calculate the object's location in

3D space based on the description of the

information obtained from individual images of the

object. The algorithm finds a common point based

on triangulation. It then calculates the individual

positions of the sensing device around the scanned

object. By subsequent calculation of the point cloud,

each point gets its own x, y, z coordinate and thus

defines the basic information about the position,

size, and geometry of the object located in 3D

space, [12]. The following Figure 1 shows the basic

point cloud and the position of individual points in

3D space. Figure 1 shows the basic shape of the

reconstructed object, and at this stage of the process,

it is ready for further processing into a 3D model.

Fig. 1: The basic principle of the photogrammetry

method.

2.1.2 LiDar Technology

This technology has been used since the 1960s.

LiDar (Light Detection and Ranging) is a standard

sensor nowadays. However, not all capturing

devices have it yet. Only the latest Apple devices

use LiDar technology, which is available to the

average user today. As the technology has advanced

over the years, LiDar has been used by many other

experts and scientists in agriculture, archaeology,

automation, mapping, and scientific fields. Use of

this technology, it is possible to determine the

distance between objects in space. LiDar also

enables a 3D representation of space and objects in

it, [13], [14].

In the experiment presented in this manuscript,

the emphasis is on the specific object and the quality

of the generated actual 3D model.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2023.20.8

Irena Drofova, Milan Adamek,

Pavel Stoklasek, Martin Ficek,

Jan Valouch

E-ISSN: 2224-3402

Volume 20, 2023

Fig. 2: The 3D object visualization using LiDar

technology.

Figure 2 shows the experimental 3DFS imaging

of an object by LiDar technology. This method of

displaying scanned data is described in more detail

in the chapter presenting the experimental results of

selected 3DFS methods.

The obtained image data from the sensing

mentioned above devices were subsequently

processed in the professional 3D modeling software

Agisoft Metashape Professional, [15]. The object

was photographed indoors with standard cold light

discharge lighting. Figure 2 shows the shine of the

wooden material on which the object was

photographed. In this case, however, the scanned

material is of a larger display of the metal part. The

degree of its shine is due to the lighting of the space

and its surroundings.

Approaches for processing the materials of a real

object into a 3D model currently represent

challenges in image processing. Especially since

each material has different surface properties, these

properties can also be seen in the environment

where the object artifacts are manipulated and

behaved. In the new case studies, the procedures

involve the use of CAD 3D models, which can lead

to incorrect simulation results. New alternative

methods are thus created to generate the most

accurate 3D models with retopological procedures

to simplify the model for further processing. These

procedures find application in digitization and 3D

modeling of large objects, [16].

2.1.3 Capture an Image for 3D Printing an

Object

The method for capturing and processing the image

for 3D printing is different in this experiment, [17].

A 3D scanner Atos Triplescan II 5M / GOM was

used to obtain a virtual 3D model of the object.

Figure 3 shows the resulting 3D model of the object.

Fig. 3: The 3D object model intended for 3D

printing.

In this case, the 3D model of the object is

scanned without the ability to define the object's

colors. That is a 3D model intended for 3D printing;

there is no information about the primary colors of

an object.

3 Experimental Results of 3DFS

In this experiment, a small object was chosen in the

form of a spent charge. The deformation of its shape

characterizes this object due to shooting. The

object's material is plastic with a black description

and metal with an engraved marking. In the 3D

modeling process, the influence of the shine from

the metallic material on the process display is

assumed. This effect is due to the cool discharge

light in the room. The object and the effect of light

can be seen in Figure 4.

Fig. 4: An object intended for 3DFS display.

3.1 3DFS Imaging Process

The hands-on experiment that this manuscript

describes is based on the sensing of several types of

sensing devices. A 3D scanner, a mobile phone

camera, and a tablet with a LiDar sensor are used.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2023.20.8

Irena Drofova, Milan Adamek,

Pavel Stoklasek, Martin Ficek,

Jan Valouch

E-ISSN: 2224-3402

Volume 20, 2023

Fig. 5: The 3DFS basic methods image processing.

All the mentioned types of sensing devices

require a different approach for capturing turnover

data and image processing for the final output. The

possibilities in the field of 3DFS are very extensive.

Figure 5 describes the basic procedures of image

data processing by individual input devices for

obtaining image data.Figure 5 describes the image

processing process of individual scanning devices.

Images captured by a mobile phone and processed

using photogrammetry have six basic steps. The first

step is to generate a point cloud from captured

photos and then process it into a 3D dense cloud

model. After this image processing, a 3D network

model can be created. This model provides the basis

for generating a solid, shadow, or textured 3D

model. The type of 3D model is chosen according to

the desired final output. The 3D model created in

this way can be transformed into a virtual reality

environment or for 3D printing.

By using a scanning device with a LiDar sensor,

some image processing steps can be omitted.

Although it is possible to generate point clouds in

some applications, in the case of the application

used in this study, a 3D polygonal mesh was

generated from a direct textured 3D model. The next

step is to generate a 3D model for the desired

output. This procedure significantly reduces the

time for creating a 3D model of an object without

using high-performance computing equipment.

The third type of 3D model discussed in this text

is an object model created by a 3D scanner to apply

the output to a 3D object by 3D print. This

procedure is swift. That is practically a direct route.

The disadvantage is the generated 3D model without

textures and colors, as seen in Figures 3 and 8. This

problem can be solved by creating a custom color

texture model to the 3D model if the goal is a 3D

digital model applied in a digital, online, or virtual

environment. A 3D model created by 3D printing

can then be created manually with a colored texture.

However, this procedure is time-consuming and

requires skill.

In the next chapter, individual 3D models are

visually compared.

3.2 The 3D Forensic Imaging Process

Visualizations of individual types of 3D models and

final models with texture are in the following

images.

3.2.1 The 3D Model by Photogrammetry

Method

A mobile phone with a camera, which was used to

create a 3D model using the photogrammetry

method, is a commonly available device. A

significant advantage is the speed of use in the case

of photography. It is possible that taking pictures in

the dark can be difficult when most current cameras

on mobile phones do not have a flash of sufficient

quality. In this experiment, 47 images were used for

the 3D modeling of the object using the

photogrammetry method. Figure 6 shows the 3D

processing of an object image into outputs obtained

by photogrammetry for forensic investigation.

a) b) c) d)

e) f) g) h)

Fig. 6: 3DFS model of the object by

photogrammetry method: a) 3D model of the

primary point cloud, b) addition of points in the

form of a dense cloud model, c) 3D object model

wireframe, d) resulting 3D model with material

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2023.20.8

Irena Drofova, Milan Adamek,

Pavel Stoklasek, Martin Ficek,

Jan Valouch

E-ISSN: 2224-3402

Volume 20, 2023

texture, e)-h) detail the 3D model of the metal part

of the object in the same image processing

procedure.

As can be seen from Figure 6, the

photogrammetry method is suitable for modeling

this object. It is more time-consuming for the entire

process and the quality requirements of all images

from which the model is created. However, the great

advantage of this method appears to be the

possibility of editing image information already in

the first two stages of image processing, Figure 6 a),

b), e), f). When this method is applied correctly, the

3D model of the metal part of the object can be

successfully modeled, as seen in Figure 6 h).

3.2.2 The 3D Model using by LiDar Sensor

A tablet with a LiDar function provides a primary

image. That function allows us to work with both

video formats and photos; both outputs can also be

combined. The photos taken can be modeled using

the photogrammetry method. It is also possible to

determine the captured video into individual

sequences and photographs. In this experiment, the

entire image was acquired in one step.

The freely available Scaniverse Pro application

was used in this experiment. This application

enables the direct transformation of the scanned

object into a 3D wireframe model. However, it does

not allow working with the model in the form of

point clouds. The scanned model was exported to

the professional modeling software Agisoft

Metashape Professional to compare the created 3D

object with the 3D object created by the

photogrammetry method.

a) b) c)

Fig. 7: The 3DFS model of the object by LiDar

sensor: a) 3D wireframe model, b) 3D model with

material texture, e) and d) detail of the 3D model of

the metal part of the object.

From the set of images 6 and 7, it can be seen

that different sensors of the sensing devices also

provide a different result in the final display of the

3D model of the object. Differences are also visible

in the polygonal network density (wireframe)

connecting individual points in the object. In Figure

7 a), it is seen that the 3D wireframe model contains

a significantly higher density of polygonal mesh

structure. With this phenomenon, the requirements

for the capacity of computing technology in image

processing and final display in a virtual environment

also grow.

3.2.3 The 3D Object Model for 3D Print

This process of capturing the image for the final 3D

model is different from the previous two cases of

3D modeling. A 3D scanner scans an object with

higher accuracy but without information about the

color and structure of the object. Currently, 3D

printing cannot work with this attribute. As Figures

3 and 8 shows, the shape of the 3D model for the

final print output contains a minimum of defects.

However, the difference in the materials of the

object is not noticeable.

Fig. 8: Detail of the 3D object model intended for

3D printing.

The advantage of the 3D printed model is the

detailed capture of the deformation of the cartridge

case material. The metal part of the object and its

relief are also clearly visible. In a forensic

investigation, these models can simulate the original

object to find another fact when analyzing the

evidence. In this digital 3D model, the deformation

of the material due to mechanical damage to the

object can be marked in more detail during further

image processing. These procedures can be applied

to the digital archiving of evidence. The

disadvantage is losing image information about

color, material, and texture.

4 Discussion

Modern procedures in the area of forensic science,

but also in many other scientific fields, require the

flexibility of approach and solving given problems.

Especially in the progressive, dynamic pace of

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2023.20.8

Irena Drofova, Milan Adamek,

Pavel Stoklasek, Martin Ficek,

Jan Valouch

E-ISSN: 2224-3402

Volume 20, 2023

digitization and work with acquired information and

data.

The details that lead to a successful solution are

often crucial in criminal investigations, trials, and

expert opinions. Modern forensic art today works

with a wide variety of tools and procedures. A 3D

forensic science (3DFS) currently has a rather

complementary character. The created 3D models

must be understood as a secondary tool designed to

present and demonstrate given facts. The created

models cannot be perceived as a primary means of

evidence because they are a virtual representation of

the natural environment, object, or evidence. These

models do not have a physical character, and neither

do 3D printing reproductions of the object are not

objective facts. Nevertheless, this discipline can

significantly help in an investigation or a criminal

case.

In the experiment described in this manuscript,

there are several procedures for quickly and

efficiently creating a 3D model of an object. The

applied methods in this work can be considered fast

in securing evidence at the crime scene.

Nevertheless, the appropriate choice of the 3DFS

method depends on the desired output. Therefore, it

is necessary to approach them individually

depending on the purpose of their use. The choice of

the capture device, image processing, and final

output depends on the environment where the image

data is captured.

The object that is the subject of this experiment

is made of two materials and structures. As seen in

the set of figures 6 and 7, there is a material with a

high degree of variable physical-optical properties

in this subject. Generally, difficult materials and

structures for 3D modeling are, for example, metals,

glass, liquids, and gels. These are materials with

variable properties due to the influence of light,

temperature, weather conditions, and other

phenomena affecting the objects' shape, structure,

and material.

5 Conclusion

In conclusion, it can be stated that optimizing the

entire process of creating a 3D model or

environment is not simple and hides many

challenges for further research into 3D imaging of

natural objects or environments, especially

concerning the large quantities of problems that

occur in the entire image processing process. There

are significant potential and challenges for the

future in the field of forensic science hidden in this

issue. One can consider a very high degree of

application of machine learning within the

framework of image segmentation and other

computer and color vision methods.

Applying and using 3DFS in crime scene

investigations makes it possible to work with

evidence on a completely different scale. As a

complementary tool, 3DFS presents new challenges

in forensics. Simultaneously with the progressive

development of new technologies and devices, the

process of creating 3D models of objects or

environments can also be significantly shortened.

This trend is currently represented by commercial

devices that include the LiDar sensor. These devices

are affordable, easily portable, and easy to operate.

The scanning devices contain a high-quality camera

and a scanning sensor. Image acquisition and

processing are faster and easier than using classic

single-lens reflex cameras (DSLR) or mobile

devices without LiDar sensors. When reconstructing

an image from these devices using the

photogrammetry method, not only a large capacity

of external storage and high performance of

computing devices but also a significant time fund

for 3D reconstruction is required. Mobile devices

with the LiDar function represent a technology

where it is possible to eliminate some of the steps of

the image processing process and export the

resulting 3D model. A significant advantage is also

that it takes less time. However, it is also necessary

to consider the resulting quality of the 3D model for

the desired output. Therefore, further image

processing on a powerful computing device is not

entirely omitted. However, that technology fulfills

everything required by the photogrammetry method.

And for reconstructing a 3D object for output to a

3D printer too.

It can be stated that a device with a LiDar

sensor can, in some cases, replace entirely both

classic cameras, a mobile phone, and the mentioned

3D scanner. Nevertheless, this technology presents

additional challenges in sensor research, sensing

devices, and image processing. All these aspects

will be considered in further research in the 3D

reconstruction of objects in the field of 3DFS.

Attention will be focused mainly on working with

point clouds and 3D reconstruction of more

complex natural objects. Here, emphasis will be

placed on the processing of the object's realistic

colors, materials, and textures and its transformation

into a 3D and VR environment for use in forensic

science.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2023.20.8

Irena Drofova, Milan Adamek,

Pavel Stoklasek, Martin Ficek,

Jan Valouch

E-ISSN: 2224-3402

Volume 20, 2023

References:

[1] Bialkova S., Van Gisbergen M.S., When

sound modulates vision: VR applications for

art and entertainment, IEEE 3rd Workshop on

Everyday Virtual Reality (WEVR), 2017, pp.

1-6, doi: 10.1109/WEVR.2017.7957714.

[2] Middleton W., Shu Q., Ludwig F.,

Representing living architecture through

skeleton reconstruction from point clouds,

Scientific Reports, 12., 2022,

doi:10.1038/s41598-022-05194-y.

[3] Gultepe, E., Edward. Conturo, T., Makrehchi,

M., Predicting and grouping digitized

paintings by style using unsupervised feature

learning, Journal of Cultural Heritage 31, 13–

23, 2018 doi.:10.1016/j.culher.2017.11.008.

[4] Selivanova and col., 3D visualization of

human body internal structures surface during

stereo-endoscopic operations using computer

vision techniques, Przegląd

Elektrotechniczny, ISSN 0033-2097, 97, nr. 9,

30-33, 2021, doi:10.15199/48.2021.09.06.

[5] Huskova V., Adamova V., Boc K., Use of 3D

Screening Technology Traks in the Process

Criminal Identification (SK), 2021, Young

Science Conference 2021, Tomas Bata

University in Zlin, Faculty of Applied

Informatics, Trilobit, 2021, [online].

[6] Min Zhang, Forensic Imaging: A Powerful

Tool in Modern Forensic

Investigation, Forensic Sciences Research,

Vol. 7, Issue 3, pp. 385–392,

2022, doi.:10.1080/20961790.2021.2008705.

[7] Errickson, D., Fawcett, H., Thompson, T.J.U.,

Campbell, A., The effect of different imaging

techniques for the visualisation of evidence in

court on jury comprehension, International

Journal of Legal Medicine 134, 2020,

doi:10.1007/s00414-019-02221-y.

[8] Carew R. M., James French J., Morgan R. M.,

3D forensic science: A new field integrating

3D imaging and 3D printing in crime

reconstruction, Forensic Science

International: Synergy, Vol. 3, 2021, ISSN

2589-871X.

[9] Wilkinson C., Neave R., Skull re-assembly

and the implications for forensic facial

reconstruction, Science & Justice, Vol. 41,

Issue 3, p. 233-234, ISSN 1355-0306, 2001,

doi.:10.1016/S1355-0306(01)71902-8.

[10] Yamasaki T., Toba S., Sanders S. P., Carreon

Ch. K.,Perfusion-distention fixation of heart

specimens: a key step in immortalizing heart

specimens for wax infiltration and generating

3D imaging data sets for reconstruction and

printed 3D models, Cardiovascular

Pathology, Vol. 58, 107404, ISSN 1054-8807,

2022, doi.org/10.1016/j.carpath.2021.107404.

[11] Drofova I., Adamek M., Analysis of the

structure material of the bronze object in 3D

models point cloud, Przegląd

Elektrotechniczny, 98(3), pp. 98-101, 2022,

doi.:10.15199/48.2022.03.22.

[12] Pavelka K., Fotogrammetrie 1, České vysoké

učení technické v Praze, 2009, ISBN 978-80-

01-04249-6.

[13] Pramod K., Akshay M. C., LiDar Technology,

International Journal for Research in Applied

Science & Engineering Technology, Vol. 10,

Issue V., ISSN: 2321-9653, 2022,

doi:10.22214/ijraset.2022.43007.

[14] Mikita, T., Krausková, D., Hrůza, P., Cibulka,

M., Patočka, Z., Forest Road Wearing Course

Damage Assessment Possibilities with

Different Types of Laser Scanning Methods

including New iPhone LiDAR Scanning

Apps, Forests, 2022, doi.:10.3390/f13111763.

[15] Agisoft, 2023, https://www.agisoft.com/

[16] Gonizzi Barsanti S., Guagliano M., Rossi A.,

3D Reality-Based Survey and Retopology for

Structural Analysis of Cultural

Heritage, Sensors, 2022,

doi.:10.3390/s22249593.

[17] Pavlovich, S., Same source profiling of 3D

printed firearms using deposition striae: a

discussion, Australian Journal of Forensic

Sciences, 2022,

DOI:10.1080/00450618.2022.2131910.

Sources of Funding for Research Presented in a

Scientific Article or Scientific Article Itself

This research was based on the support of the

International Grand Agency of Tomas Bata

University in Zlín, No. IGA/CebiaTech/2022/004,

and the Department of Security Engineering,

Faculty of Applied Informatics.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2023.20.8

Irena Drofova, Milan Adamek,

Pavel Stoklasek, Martin Ficek,

Jan Valouch

E-ISSN: 2224-3402

Volume 20, 2023

Contribution of Individual Authors to the

Creation of a Scientific Article (Ghostwriting

Policy)

The authors equally contributed in the present

research, at all stages from the formulation of the

problem to the final findings and solution.

Conflict of Interest

The authors have no conflicts of interest to declare

that are relevant to the content of this article.

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the

Creative Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.en

_US